Research On Image Recognition Method Of Wheat Rust Based On Open CV

Wheat is an important staple food for human beings.But rust is one of the diseases that can wreak havoc on wheat as it grows.In this paper,the plant protection uav was used to collect images of wheat in the middle growth stage,and the characteristics of wheat rust images were extracted through the Open CV visual library.The abundant visual processing functions were used to identify the rust in wheat leaves and stems.In the later stage,plant protection drones can be used to spray targeted drugs on wheat rust areas,so as to achieve the purpose of precise drug use and precise management of wheat rust.The main content of this paper is to identify wheat rust based on Open CV,and identify the characteristics of leaf and stem rust of wheat respectively.The main research contents and conclusions are as follows:The image denoising methods based on spatial domain include mean filtering,median filtering and wiener filtering.The image denoising methods based on frequency domain include Fourier transform,Gabor transform and wavelet transform.The image denoising methods based on morphology include corrosion and expansion.After the same wheat rust image denoising method using the above treatment,combined with the peak signal to noise ratio and mean square error is analyzed,the result of the denoising results show:the morphological open operation for best denoising method,its peak ratio 38.0556,the mean square error is 10.1746,among the peak signal to noise ratio in the denoising method of maximum and minimum mean square error(mse).2.Wheat rust image segmentation(the foreground of segmentation is the disease part of wheat)the wheat disease image is segmented by histogram threshold segmentation method,regional growth algorithm and k-means segmentation algorithm,and the segmentation image is evaluated based on the maximum entropy.The results show that k-means segmentation algorithm is the best.3.Extraction of image features of wheat rust color features,texture features and shape features were used to extract the disease part of the wheat image.According to the visual effect,the best feature extraction method was based on the S component of HSV color space.4.Finally,support vector machine(SVM)was used to process the wheat disease image,and polynomial,radial basis and Sigmoid kernel functions were used to extract the characteristics of the wheat disease image,and the maximum entropy results showed that the SVM with polynomial kernel function had the best extraction effect on the wheat disease parts.The above research results provide a summary of identification methods for the study of wheat disease images,and provide a new test method for wheat disease images based on color space,which lays a foundation for the recognition of wheat disease images based on Open CV,and is conducive to the research on identification and detection of crop diseases.